In reference to David's comments at the bottom of this, let me clarify
by modifying my statement to say that anything consistently outside of
the range of .3-.8 volts needs to be corrected.

As to the notion that the ECU adds fuel pulse until the other trigger
point is reached, I don't believe this is the case. Darrel would
probably be a better source when he returns.  In my observations it
appears that the ECU has a sampling interval. In other words when a
trigger point is reach, an injector pulse adjustment is made and then an
interval before another adjustment is made.  I say this because using
both a LED type O2 tester, and a digital voltmeter during driving the
blink rate of the LED was noticeably different and the voltage at times
was with the .3 -.8 volt range for a number of seconds before the ECU
made an adjustment.

I surmised the ECU doesn't keep adding or taking away fuel until the
other trigger point is reached  but rather waits until O2 voltage again
exceeds the parameter, and then the cycle is repeated.

When troubleshooting the Fuel Injection system, an averaging Voltmeter
is supposed to be used to see what the average O2 voltage is. When this
type of averaging Voltmeter, with a memory is used, the lower, average
and upper volt output of the O2 sensor, in closed loop of course, are
obtainable.  From my testing on my van the average low was .3 and high
.8.

I've found that it is easier for me adjust my fuel injection mixture by
disconnect the O2 sensor to put the ECU in open loop and then  monitor
the O2 readings while driving.  Using this method I can adjust spring
tension , or in my case fuel pressure or wiper arm position, so that the
ECU maintains the fuel mixture around a target voltage.  I target at .7
volts.  Then I reconnect my O2 sensor and consider that part of the
injection system adjusted.

Sincerely,

Mark Keller
91 Carat

and should be in the .5 volt range.  Anything consistently above .8 or
below .3 volts needs to be corrected.

I'd be interested in expert opinion on this.  The measurement range of
the
lambda sensor is so tiny that it really only indicates "perfect," "too
rich," and "too lean" without giving any useful idea of how much.  You
can
get lambda sensors with an extended range, but they cost many hundreds
of
dollars each, and wouldn't work properly with the ECU anyway.  But if
you
were rich you could add one as a monitoring device to read out on a
meter
or LED bar graph.

When the system is really tuned the O2 sensor will not crossover as
rapidly as a system that is too rich,

This is wrong.  The ECU treats the lambda sensor as having only two
states,
too rich and too lean.  Whichever way it finds, it kicks the mixture
until
it goes the other way, then kicks it back, repeat forever.  This goes on

many times a second at high rpm, and on an oscilloscope you could watch
as
it happens.  If you stick a regular digital voltmeter on it you'll
simply
see a number that jumps around seemingly randomly, because the voltmeter

integrates over a noticeable chunk of a second to get a reading.

david

David Beierl - Providence, RI
  http://pws.prserv.net/synergy/Vanagon/
'84 Westy "Dutiful Passage"
'85 GL "Poor Relation"

------------------------------